Topic: Currency Backed by Computing Power (Read 1075 times)
You've given me a lot to think about. When I come up with more ideas, I'll put them up.
I'm not necessarily thinking of this as a currency first...but as a way for someone to bank computing power. I don't need all of my GFX card's power all the time, but there are occasions where I need much more than it can provide -- same with my CPU. It would be nice if I could store that power (as a coin) and use it when I need it most.
I was actually just thinking that perhaps, when the demand was low for computational power, the network could be securing a chain -- or perhaps up to x% or total power will always be creating the chain. As for verifying work, you would have to run the "official miner" which would check itself against the other miners on the network, or you could have two separate miners work on the problem, and compare answers...those are very good questions, keep them coming, please, as we work this out.
Well I think you have it backwards. The blockchain must ALWAYS be protected with massive amounts of computing power. The only way I would see such a system as useful is that the protocol allows up to some x% of network power to be used for private work. x should be some small %. Say <20% that way 80%-100% of network power is securing the block chain. If there is insufficient network power available for private work transactions will either be queued or people could bid higher to gain priority (i.e. spend more than 1 xCoin to get 1 unit of computational work).
Granted I have no idea if this is even feasible but the highest priority should always be to protect the block chain. If the block chain is compromised the currency is worthless and thus the idea is worthless.
"I don't think this could implement a blockchain, persay -- as the current hashpower is required to maintain one. The coins would have to be secured and verified some other way. "
This is a non trivial issue. I mean if you simply explain it away and then don't come up with a solution it is DOA. The system needs to be secured by something and currently every distributed implementation is secured via massive computational power.
That doesn't mean coins couldn't be redeemed for time on the network. Bitcoin doesn't currently implement that but unless everyone "redeemed" a lot of coins at once there is no reason the network can't do both. In essence you need computing power. You redeem one xCoin via a transaction and it gets allocated via the network to a node. The node does the work and gets x. You get the results. The issue them becomes avoiding counterfit work.
What if I do your "work" and simply return all zeroes. There are potential solutions but those solutions add more problems. It is an interesting idea however starting with "the coins would need to be secured another way is usless unless you have another way".
A more useful academic exercise is:
1) I know distributed computing power can secure a block chain
2) I know some % of that can be siphoned off to do other works (with compensation) without significantly increasing the risk to blockchain.
3) Can I somehow come up with a system which uses 1 & 2 to fix a coin to some amount of usable computing power?
This is a non trivial issue. I mean if you simply explain it away and then don't come up with a solution it is DOA. The system needs to be secured by something and currently every distributed implementation is secured via massive computational power.
That doesn't mean coins couldn't be redeemed for time on the network. Bitcoin doesn't currently implement that but unless everyone "redeemed" a lot of coins at once there is no reason the network can't do both. In essence you need computing power. You redeem one xCoin via a transaction and it gets allocated via the network to a node. The node does the work and gets x. You get the results. The issue them becomes avoiding counterfit work.
What if I do your "work" and simply return all zeroes. There are potential solutions but those solutions add more problems. It is an interesting idea however starting with "the coins would need to be secured another way is usless unless you have another way".
A more useful academic exercise is:
1) I know distributed computing power can secure a block chain
2) I know some % of that can be siphoned off to do other works (with compensation) without significantly increasing the risk to blockchain.
3) Can I somehow come up with a system which uses 1 & 2 to fix a coin to some amount of usable computing power?
I was actually just thinking that perhaps, when the demand was low for computational power, the network could be securing a chain -- or perhaps up to x% or total power will always be creating the chain. As for verifying work, you would have to run the "official miner" which would check itself against the other miners on the network, or you could have two separate miners work on the problem, and compare answers...those are very good questions, keep them coming, please, as we work this out.
USD, for example, started as backed by something (gold) and are now backed by millions of users and the US military.
USD is backed by taxation system. You need USD or you will go to jail.Yes, and they send you to jail with the military might they possess.
Sounds really interesting, but the potential for malicious use of the rented power would surely be huge?
Yes, it could be used for evil. But so can lots of things -- I don't think that's a reason to not pursue it.
This idea is something I'm still working on -- any suggestions are welcome, criticisms etc.
Sounds really interesting, but the potential for malicious use of the rented power would surely be huge?
USD, for example, started as backed by something (gold) and are now backed by millions of users and the US military.
USD is backed by taxation system. You need USD or you will go to jail.
There are some interesting ideas.
USD, for example, started as backed by something (gold) and are now backed by millions of users and the US military.
USD is backed by taxation system. You need USD or you will go to jail.
"I don't think this could implement a blockchain, persay -- as the current hashpower is required to maintain one. The coins would have to be secured and verified some other way. "
This is a non trivial issue. I mean if you simply explain it away and then don't come up with a solution it is DOA. The system needs to be secured by something and currently every distributed implementation is secured via massive computational power.
That doesn't mean coins couldn't be redeemed for time on the network. Bitcoin doesn't currently implement that but unless everyone "redeemed" a lot of coins at once there is no reason the network can't do both. In essence you need computing power. You redeem one xCoin via a transaction and it gets allocated via the network to a node. The node does the work and gets x. You get the results. The issue them becomes avoiding counterfit work.
What if I do your "work" and simply return all zeroes. There are potential solutions but those solutions add more problems. It is an interesting idea however starting with "the coins would need to be secured another way is usless unless you have another way".
A more useful academic exercise is:
1) I know distributed computing power can secure a block chain
2) I know some % of that can be siphoned off to do other works (with compensation) without significantly increasing the risk to blockchain.
3) Can I somehow come up with a system which uses 1 & 2 to fix a coin to some amount of usable computing power?
This is a non trivial issue. I mean if you simply explain it away and then don't come up with a solution it is DOA. The system needs to be secured by something and currently every distributed implementation is secured via massive computational power.
That doesn't mean coins couldn't be redeemed for time on the network. Bitcoin doesn't currently implement that but unless everyone "redeemed" a lot of coins at once there is no reason the network can't do both. In essence you need computing power. You redeem one xCoin via a transaction and it gets allocated via the network to a node. The node does the work and gets x. You get the results. The issue them becomes avoiding counterfit work.
What if I do your "work" and simply return all zeroes. There are potential solutions but those solutions add more problems. It is an interesting idea however starting with "the coins would need to be secured another way is usless unless you have another way".
A more useful academic exercise is:
1) I know distributed computing power can secure a block chain
2) I know some % of that can be siphoned off to do other works (with compensation) without significantly increasing the risk to blockchain.
3) Can I somehow come up with a system which uses 1 & 2 to fix a coin to some amount of usable computing power?
Hello everyone, I was thinking about Bitcoin the other day, and a problem frequently stated by detractors : that BTC is not backed by anything, except a few thousand users hopes and dreams. Yes, I understand it is a "fiat" currency, but the primary difference is other fiat currencies -- USD, for example, started as backed by something (gold) and are now backed by millions of users and the US military. So I got to thinking, what resource do we have that could make a digital currency more "real." I briefly toyed around with the idea of a coin backed by programming ability, but that's a resource most of us already exchange for USD. But we have gigahashes of computing power, devoted to pushing coins around. Is there a way to use that for something more?
What I'm thinking is a coin, backed by our hashpower. Coins would be directly and immediately exchangeable for time on the network -- think about it. I want to render a scene in Maya, and it would normally take me hours on my system at home. But if I was able to use the entire BTC network, it would be a manner of seconds -- I would pay in "timecoins" or whatever, and be on my way. In return for letting me use your hash power, the network would award you x amount of "timecoins", which you would be free to redeem for computing power at a later date, or sell to a company/individual who needs to run some code.
I don't think this could implement a blockchain, persay -- as the current hashpower is required to maintain one. The coins would have to be secured and verified some other way. The other issue is of course making this huge network work with current software -- there would probably have to be a driver so the network would show up as a graphics adapter. Other details would be how to measure time on the network? Clock cycles? I've also thought about having two networks, one for CPUs, one for GPUs, depending on the work you need to do.
Maybe the system could even be used to provide an OnLive style of gaming -- where you're essentially playing a stream that's hosted on the network.
I am a CS major, and this is a little outside of my expertise (have done hardly any driver work). I think it's an interesting concept, at least, and possibly worth exploring further. We need to make a digital currency that is directly redeemable for something if we want it to catch on in a timely manner.
What I'm thinking is a coin, backed by our hashpower. Coins would be directly and immediately exchangeable for time on the network -- think about it. I want to render a scene in Maya, and it would normally take me hours on my system at home. But if I was able to use the entire BTC network, it would be a manner of seconds -- I would pay in "timecoins" or whatever, and be on my way. In return for letting me use your hash power, the network would award you x amount of "timecoins", which you would be free to redeem for computing power at a later date, or sell to a company/individual who needs to run some code.
I don't think this could implement a blockchain, persay -- as the current hashpower is required to maintain one. The coins would have to be secured and verified some other way. The other issue is of course making this huge network work with current software -- there would probably have to be a driver so the network would show up as a graphics adapter. Other details would be how to measure time on the network? Clock cycles? I've also thought about having two networks, one for CPUs, one for GPUs, depending on the work you need to do.
Maybe the system could even be used to provide an OnLive style of gaming -- where you're essentially playing a stream that's hosted on the network.
I am a CS major, and this is a little outside of my expertise (have done hardly any driver work). I think it's an interesting concept, at least, and possibly worth exploring further. We need to make a digital currency that is directly redeemable for something if we want it to catch on in a timely manner.
Jump to: